Eric Dumazet [Wed, 27 Apr 2011 22:56:07 +0000 (22:56 +0000)]
net: dont hold rtnl mutex during netlink dump callbacks
Four years ago, Patrick made a change to hold rtnl mutex during netlink
dump callbacks.
I believe it was a wrong move. This slows down concurrent dumps, making
good old /proc/net/ files faster than rtnetlink in some situations.
This occurred to me because one "ip link show dev ..." was _very_ slow
on a workload adding/removing network devices in background.
All dump callbacks are able to use RCU locking now, so this patch does
roughly a revert of commits :
1c2d670f366 : [RTNETLINK]: Hold rtnl_mutex during netlink dump callbacks 6313c1e0992 : [RTNETLINK]: Remove unnecessary locking in dump callbacks
This let writers fight for rtnl mutex and readers going full speed.
It also takes care of phonet : phonet_route_get() is now called from rcu
read section. I renamed it to phonet_route_get_rcu()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: Patrick McHardy <kaber@trash.net> Cc: Remi Denis-Courmont <remi.denis-courmont@nokia.com> Acked-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 2 May 2011 21:37:45 +0000 (14:37 -0700)]
ipv4: Make sure flowi4->{saddr,daddr} are always set.
Slow path output route resolution always makes sure that
->{saddr,daddr} are set, and also if we trigger into IPSEC resolution
we initialize them as well, because xfrm_lookup() expects them to be
fully resolved.
But if we hit the fast path and flowi4->flowi4_proto is zero, we won't
do this initialization.
Therefore, move the IPSEC path initialization to the route cache
lookup fast path to make sure these are always set.
Signed-off-by: David S. Miller <davem@davemloft.net>
- gw_node_update() doesn't add a new gw_node in case of empty curr_gw.
This means that at the beginning no gw_node is added, leading to an
empty gateway list.
- gw_election() is terminating in case of curr_gw == NULL. It has to
terminate in case of curr_gw != NULL
Signed-off-by: Antonio Quartulli <ordex@autistici.org> Signed-off-by: Sven Eckelmann <sven@narfation.org>
orig_hash_find() manages rcu_lock/unlock internally and doesn't need to
be surrounded by rcu_read_lock() / rcu_read_unlock() anymore
Signed-off-by: Antonio Quartulli <ordex@autistici.org> Acked-by: Marek Lindner <lindner_marek@yahoo.de> Signed-off-by: Sven Eckelmann <sven@narfation.org>
David Decotigny [Wed, 27 Apr 2011 18:32:43 +0000 (18:32 +0000)]
net/igb/e1000/e1000e: more robust ethtool duplex/speed configuration
This makes sure that one cannot request a 99Mbps full-duplex and get a
100Mbps half-duplex configuration in return due to the way the
speed/duplex parameters are handled internally.
Tested: e1000 works Signed-off-by: David Decotigny <decot@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Decotigny [Wed, 27 Apr 2011 18:32:42 +0000 (18:32 +0000)]
tulip/de2104x: don't report different speeds depending on port type
Initial driver reported different speeds depending on the port being
used. This advertises the speed to be 10Mbps in any case, which is
what it actually is on the wire.
Signed-off-by: David Decotigny <decot@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Decotigny [Wed, 27 Apr 2011 18:32:40 +0000 (18:32 +0000)]
ethtool: cosmetic: Use ethtool ethtool_cmd_speed API
This updates the network drivers so that they don't access the
ethtool_cmd::speed field directly, but use ethtool_cmd_speed()
instead.
For most of the drivers, these changes are purely cosmetic and don't
fix any problem, such as for those 1GbE/10GbE drivers that indirectly
call their own ethtool get_settings()/mii_ethtool_gset(). The changes
are meant to enforce code consistency and provide robustness with
future larger throughputs, at the expense of a few CPU cycles for each
ethtool operation.
All drivers compiled with make allyesconfig ion x86_64 have been
updated.
Tested: make allyesconfig on x86_64 + e1000e/bnx2x work Signed-off-by: David Decotigny <decot@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Decotigny [Wed, 27 Apr 2011 18:32:39 +0000 (18:32 +0000)]
ethtool: Use full 32 bit speed range in ethtool's set_settings
This makes sure the ethtool's set_settings() callback of network
drivers don't ignore the 16 most significant bits when ethtool calls
their set_settings().
All drivers compiled with make allyesconfig on x86_64 have been
updated.
Signed-off-by: David Decotigny <decot@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Decotigny [Wed, 27 Apr 2011 18:32:38 +0000 (18:32 +0000)]
ethtool: Call ethtool's get/set_settings callbacks with cleaned data
This makes sure that when a driver calls the ethtool's
get/set_settings() callback of another driver, the data passed to it
is clean. This guarantees that speed_hi will be zeroed correctly if
the called callback doesn't explicitely set it: we are sure we don't
get a corrupted speed from the underlying driver. We also take care of
setting the cmd field appropriately (ETHTOOL_GSET/SSET).
This applies to dev_ethtool_get_settings(), which now makes sure it
sets up that ethtool command parameter correctly before passing it to
drivers. This also means that whoever calls dev_ethtool_get_settings()
does not have to clean the ethtool command parameter. This function
also becomes an exported symbol instead of an inline.
All drivers visible to make allyesconfig under x86_64 have been
updated.
Signed-off-by: David Decotigny <decot@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David Decotigny [Wed, 27 Apr 2011 18:32:37 +0000 (18:32 +0000)]
ethtool: cosmetics: enforce const-ness in ethtool_cmd_speed
The 'ep' argument of ethtool_cmd_speed is not altered: advertise it in
protoype. +Indentation fix. Also add comments to advise using the
ethtool_cmd_speed API to get/set the link speed.
Signed-off-by: David Decotigny <decot@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Sony Chacko [Thu, 28 Apr 2011 11:48:19 +0000 (11:48 +0000)]
qlcnic: Support for GBE port settings
Enable setting speed and auto negotiation parameters for GbE ports.
Hardware do not support half duplex setting currently.
o Update driver version to 5.0.17.
Signed-off-by: Sony Chacko <sony.chacko@qlogic.com> Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
qlcnic: support rcv ring configuration through ethtool
o Support ethtool command ETHTOOL_GCHANNELS and ETHTOOL_SCHANNELS.
o Number of rcv rings configuration depend upon number of msix vector.
Signed-off-by: Sucheta Chakraborty <sucheta.chakraborty@qlogic.com> Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Hutchings [Tue, 26 Apr 2011 15:25:52 +0000 (15:25 +0000)]
ipv4, ipv6, bonding: Restore control over number of peer notifications
For backward compatibility, we should retain the module parameters and
sysfs attributes to control the number of peer notifications
(gratuitous ARPs and unsolicited NAs) sent after bonding failover.
Also, it is possible for failover to take place even though the new
active slave does not have link up, and in that case the peer
notification should be deferred until it does.
Change ipv4 and ipv6 so they do not automatically send peer
notifications on bonding failover.
Change the bonding driver to send separate NETDEV_NOTIFY_PEERS
notifications when the link is up, as many times as requested. Since
it does not directly control which protocols send notifications, make
num_grat_arp and num_unsol_na aliases for a single parameter. Bump
the bonding version number and update its documentation.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: Jay Vosburgh <fubar@us.ibm.com> Acked-by: Brian Haley <brian.haley@hp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Fri, 29 Apr 2011 17:20:53 +0000 (10:20 -0700)]
bpf: depends on MODULES
module_alloc() and module_free() are available only if CONFIG_MODULES=y
Reported-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 28 Apr 2011 20:54:06 +0000 (13:54 -0700)]
l2tp: Fix inet_opt conversion.
We don't actually hold the socket lock at this point, so the
rcu_dereference_protected() isn't' correct. Thanks to Eric
Dumazet for pointing this out.
Thankfully, we're only interested in fetching the faddr value
if srr is enabled, so we can simply make this an RCU sequence
and use plain rcu_dereference().
Reported-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michał Mirosław [Thu, 28 Apr 2011 01:59:15 +0000 (11:59 +1000)]
net: ibmveth: force reconfiguring checksum settings on startup
Commit b9367bf3ee6d ("net: ibmveth: convert to hw_features") accidentally
removed call to ibmveth_set_csum_offload() in ibmveth_probe(). Put the
call back where it was, but with additional error checking provided
by ibmveth_set_features().
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl> Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
[sfr: dev -> netdev] Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
o Memory allocated in ETHTOOL_ACTIVE mode, is not getting freed. So,
in ETHTOOL_ID_INACTIVE mode, return after freeing allocated memory.
o Using set bit instead of blink_down field, as it is also required
in internal Loopback test and etc.
Signed-off-by: Sucheta Chakraborty <sucheta.chakraborty@qlogic.com> Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michał Mirosław [Fri, 22 Apr 2011 06:31:16 +0000 (06:31 +0000)]
net: fix netdev_increment_features()
Simplify and fix netdev_increment_features() to conform to what is
stated in netdevice.h comments about NETIF_F_ONE_FOR_ALL.
Include FCoE segmentation and VLAN-challedged flags in computation.
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Thu, 21 Apr 2011 09:45:37 +0000 (09:45 +0000)]
inet: add RCU protection to inet->opt
We lack proper synchronization to manipulate inet->opt ip_options
Problem is ip_make_skb() calls ip_setup_cork() and
ip_setup_cork() possibly makes a copy of ipc->opt (struct ip_options),
without any protection against another thread manipulating inet->opt.
Another thread can change inet->opt pointer and free old one under us.
Use RCU to protect inet->opt (changed to inet->inet_opt).
Instead of handling atomic refcounts, just copy ip_options when
necessary, to avoid cache line dirtying.
We cant insert an rcu_head in struct ip_options since its included in
skb->cb[], so this patch is large because I had to introduce a new
ip_options_rcu structure.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
BPF program is 144 bytes long, so native program is almost same size ;)
(000) ldh [12]
(001) jeq #0x800 jt 2 jf 8
(002) ld [26]
(003) and #0xffffff00
(004) jeq #0xc0a81400 jt 16 jf 5
(005) ld [30]
(006) and #0xffffff00
(007) jeq #0xc0a81400 jt 16 jf 17
(008) jeq #0x806 jt 10 jf 9
(009) jeq #0x8035 jt 10 jf 17
(010) ld [28]
(011) and #0xffffff00
(012) jeq #0xc0a81400 jt 16 jf 13
(013) ld [38]
(014) and #0xffffff00
(015) jeq #0xc0a81400 jt 16 jf 17
(016) ret #65535
(017) ret #0
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: Arnaldo Carvalho de Melo <acme@infradead.org> Cc: Ben Hutchings <bhutchings@solarflare.com> Cc: Hagen Paul Pfeifer <hagen@jauu.net> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 26 Apr 2011 20:57:47 +0000 (13:57 -0700)]
ipv4: Remove erroneous check in igmpv3_newpack() and igmp_send_report().
Output route resolution never returns a route with rt_src set to zero
(which is INADDR_ANY).
Even if the flow key for the output route lookup specifies INADDR_ANY
for the source address, the output route resolution chooses a real
source address to use in the final route.
This test has existed forever in igmp_send_report() and David Stevens
simply copied over the erroneous test when implementing support for
IGMPv3.
Signed-off-by: David S. Miller <davem@davemloft.net> Reviewed-by: Eric Dumazet <eric.dumazet@gmail.com>
David S. Miller [Tue, 26 Apr 2011 20:28:44 +0000 (13:28 -0700)]
ipv4: Sanitize and simplify ip_route_{connect,newports}()
These functions are used together as a unit for route resolution
during connect(). They address the chicken-and-egg problem that
exists when ports need to be allocated during connect() processing,
yet such port allocations require addressing information from the
routing code.
It's currently more heavy handed than it needs to be, and in
particular we allocate and initialize a flow object twice.
Let the callers provide the on-stack flow object. That way we only
need to initialize it once in the ip_route_connect() call.
Later, if ip_route_newports() needs to do anything, it re-uses that
flow object as-is except for the ports which it updates before the
route re-lookup.
Also, describe why this set of facilities are needed and how it works
in a big comment.
Signed-off-by: David S. Miller <davem@davemloft.net> Reviewed-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com> Cc: Realtek linux nic maintainers <nic_swsd@realtek.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Change the call to take the transport parameter and set the
cached 'dst' appropriately inside the get_dst() function calls.
This will allow us in the future to clean up source address
storage as well.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
sctp: remove useless arguments from get_saddr() call
There is no point in passing a destination address to
a get_saddr() call.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
SCTP never called xfrm_output after it's v6 route lookups so
that never really worked with ipsec. Additioanlly, we never
passed port nubmers and protocol in the flowi, so any port
based policies were never applied as well. Now that we can
fixed ipv6 routing lookup code, using ip6_dst_lookup_flow()
and pass port numbers.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The ipv6 routing lookup does give us a source address,
but instead of filling it into the dst, it's stored in
the flowi. We can use that instead of going through the
entire source address selection again.
Also the useless ->dst_saddr member of sctp_pf is removed.
And sctp_v6_dst_saddr() is removed, instead by introduce
sctp_v6_to_addr(), which can be reused to cleanup some dup
code.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Weixing Shi [Tue, 26 Apr 2011 21:36:32 +0000 (21:36 +0000)]
sctp: fix sctp to work with ipv6 source address routing
In the below test case, using the source address routing,
sctp can not work.
Node-A
1)ifconfig eth0 inet6 add 2001:1::1/64
2)ip -6 rule add from 2001:1::1 table 100 pref 100
3)ip -6 route add 2001:2::1 dev eth0 table 100
4)sctp_darn -H 2001:1::1 -P 250 -l &
Node-B
1)ifconfig eth0 inet6 add 2001:2::1/64
2)ip -6 rule add from 2001:2::1 table 100 pref 100
3)ip -6 route add 2001:1::1 dev eth0 table 100
4)sctp_darn -H 2001:2::1 -P 250 -h 2001:1::1 -p 250 -s
root cause:
Node-A and Node-B use the source address routing, and
at begining, source address will be NULL,sctp will
search the routing table by the destination address,
because using the source address routing table, and
the result dst_entry will be NULL.
solution:
walk through the bind address list to get the source
address and then lookup the routing table again to get
the correct dst_entry.
Signed-off-by: Weixing Shi <Weixing.Shi@windriver.com> Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Don Skidmore [Sat, 16 Apr 2011 05:29:14 +0000 (05:29 +0000)]
ixgbe: fix X540 ethtool loopback test.
On X540 we need to set the MACC.FLU bit to 1 in order to force the link
up before entering MAC loopback. This is only used in the ethtool loopback
test, which was failing. This patch corrects it.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Evan Swanson <evan.swanson@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Yi Zou [Sat, 9 Apr 2011 08:34:12 +0000 (08:34 +0000)]
ixgbe: do not clear FCoE DDP error status for received ABTS
The ddp->err is initialized to be 1 to make sure outstanding DDP context is
guaranteed to be invalidated when HW is not auto-invalidating it. However,
in case of receiving ABTS response for a DDPed I/O, the ddp->err was cleared,
bypassing the invalidating of the DDP context from upper protocol stack when
ixgbe_fcoe_ddp_put() is called. This bug is fixed here by updating the error
only when FCP_RSP is received.
Signed-off-by: Yi Zou <yi.zou@intel.com> Tested-by: Ross Brattain <ross.b.brattain@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Alexander Duyck [Sat, 9 Apr 2011 05:34:06 +0000 (05:34 +0000)]
ixgbe: remove ntuple display support
This change removes the ntuple display support from ixgbe. The reason for
this change is to resolve a number of issues in the way display filtering
is handled.
I plan to add support for displaying these filters via the network flow
classifier interface.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com> Tested-by: Evan Swanson <evan.swanson@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Emil Tantilov [Fri, 1 Apr 2011 08:17:19 +0000 (08:17 +0000)]
ixgbe: add support for 64k EEPROM for 82599
82599 supports up to 32k EEPROM addressing via EERD register. If we
wish to address larger EEPROM this have to be done via serial interface.
This patch adds function ixgbe_read_eeprom_82599 which selects the best
method to read the EEPROM.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Evan Swanson <evan.swanson@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Bruce Allan [Sat, 16 Apr 2011 00:34:40 +0000 (00:34 +0000)]
e1000e: implement ethtool set_phys_id
Based on a patch from Stephen Hemminger <shemminger@vyatta.com>.
The new ethtool set_phys_id takes over controlling the LED for
identifying boards. This fixes the lockout during that period.
For this device lots of extra infrastructure can also be removed by
using set_phys_id.
v2: - return blink frequency for parts that do not support blink in h/w
- add blink_led function pointers for devices that do support blink
in h/w to cleanup the test for this functionality
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Cc: Stephen Hemminger <shemminger@vyatta.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Matt Carlson [Mon, 25 Apr 2011 12:42:50 +0000 (12:42 +0000)]
tg3: Update version to 3.118
This patch updates the tg3 version to 3.118.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 25 Apr 2011 12:42:49 +0000 (12:42 +0000)]
tg3: Whitespace cleanups
This patch gets rid of some harmless whitespace errors.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 25 Apr 2011 12:42:48 +0000 (12:42 +0000)]
tg3: Add EEH support
This patch adds EEH support to the tg3 driver.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 25 Apr 2011 12:42:47 +0000 (12:42 +0000)]
tg3: Add TSO loopback test
This patch adds code to exercise the TSO portion of the device through
a phy loopback test.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Benjamin Li <benli@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 25 Apr 2011 12:42:46 +0000 (12:42 +0000)]
tg3: Organize loopback test failure flags
As more test modes are added to each loopback mode, the need to
organise the results increases. This patch groups the results by
loopback mode, and then by test mode.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Mon, 25 Apr 2011 12:42:45 +0000 (12:42 +0000)]
tg3: Fix int generation hw bug for 5719 / 5720
On the 5719 and 5720, there is a bug where the hardware will
misinterpret a status tag update and leave interrupts permanently
disabled. This patch enables a hardware fix that works around the
issue.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Reviewed-by: Benjamin Li <benli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Mon, 25 Apr 2011 20:03:02 +0000 (13:03 -0700)]
bluetooth: Fix use-before-initiailized var.
net/bluetooth/l2cap_core.c: In function ‘l2cap_recv_frame’:
net/bluetooth/l2cap_core.c:3612:15: warning: ‘sk’ may be used uninitialized in this function
net/bluetooth/l2cap_core.c:3612:15: note: ‘sk’ was declared here
Actually the problem is in the inline function l2cap_data_channel(), we
branch to the label 'done' which tests 'sk' before we set it to anything.
Initialize it to NULL to fix this.
Signed-off-by: David S. Miller <davem@davemloft.net>
Same WRB entry was being reused over different iterations of a
loop while issuing non-embedded IOCTL requests.Fixed couple of minor bugs
in this path as well.
Re-factored code to alloc/free memory for DMA outside of loop
Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com> Signed-off-by: David S. Miller <davem@davemloft.net>
bonding: move processing of recv handlers into handle_frame()
Since now when bonding uses rx_handler, all traffic going into bond
device goes thru bond_handle_frame. So there's no need to go back into
bonding code later via ptype handlers. This patch converts
original ptype handlers into "bonding receive probes". These functions
are called from bond_handle_frame and they are registered per-mode.
Note that vlan packets are also handled because they are always untagged
thanks to vlan_untag()
Note that this also allows arpmon for eth-bond-bridge-vlan topology.
Signed-off-by: Jiri Pirko <jpirko@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Michał Mirosław [Thu, 21 Apr 2011 12:42:15 +0000 (12:42 +0000)]
net: make WARN_ON in dev_disable_lro() useful
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The oops happens in dst_metrics_write_ptr()
include/net/dst.h:124: return dst->ops->cow_metrics(dst, p);
dst->ops->cow_metrics is NULL and causes the oops.
Provide cow_metrics() methods, like we did in commit 214f45c91bb
(net: provide default_advmss() methods to blackhole dst_ops)
Signed-off-by: Held Bernhard <berny156@gmx.de> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The default maximum transmit length for NCM USB frames should be so
that a short packet happens at the end if the device supports a length
greater than the defined maximum. This is achieved by adding 4 bytes
to the maximum length so that the existing logic can fit a short
packet there.
Signed-off-by: Hans Petter Selasky <hselasky@c2i.net> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sun, 24 Apr 2011 17:54:56 +0000 (10:54 -0700)]
net: Remove __KERNEL__ cpp checks from include/net
These header files are never installed to user consumption, so any
__KERNEL__ cpp checks are superfluous.
Projects should also not copy these files into their userland utility
sources and try to use them there. If they insist on doing so, the
onus is on them to sanitize the headers as needed.
Signed-off-by: David S. Miller <davem@davemloft.net>
François Romieu [Sun, 24 Apr 2011 15:38:48 +0000 (17:38 +0200)]
r8169: don't request firmware when there's no userspace.
The firmware is cached during the first successfull call to open() and
released once the network device is unregistered. The driver uses the
cached firmware between open() and unregister_netdev().
So far the firmware is optional : a failure to load the firmware does
not prevent open() to success. It is thus necessary to 1) unregister
all 816x / 810[23] devices and 2) force a driver probe to issue a new
firmware load.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com> Fixed-by: Ciprian Docan <docan@eden.rutgers.edu> Cc: Realtek linux nic maintainers <nic_swsd@realtek.com>
Fixed packets parameters for FW in UDP checksum offload flow.
Do not dereference TCP headers on non TCP frames. Reported-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Neil Horman [Fri, 22 Apr 2011 08:10:59 +0000 (08:10 +0000)]
netconsole: fix deadlock when removing net driver that netconsole is using (v2)
A deadlock was reported to me recently that occured when netconsole was being
used in a virtual guest. If the virtio_net driver was removed while netconsole
was setup to use an interface that was driven by that driver, the guest
deadlocked. No backtrace was provided because netconsole was the only console
configured, but it became clear pretty quickly what the problem was. In
netconsole_netdev_event, if we get an unregister event, we call
__netpoll_cleanup with the target_list_lock held and irqs disabled.
__netpoll_cleanup can, if pending netpoll packets are waiting call
cancel_delayed_work_sync, which is a sleeping path. the might_sleep call in
that path gets triggered, causing a console warning to be issued. The
netconsole write handler of course tries to take the target_list_lock again,
which we already hold, causing deadlock.
The fix is pretty striaghtforward. Simply drop the target_list_lock and
re-enable irqs prior to calling __netpoll_cleanup, the re-acquire the lock, and
restart the loop. Confirmed by myself to fix the problem reported.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com> CC: "David S. Miller" <davem@davemloft.net> Signed-off-by: David S. Miller <davem@davemloft.net>
The commit was meant to support transport bridging, and specifically
virtual machines bridged to an ethernet interface connected to a
switch port wiht 802.1x enabled.
But this isn't the way to do it, it breaks too many other things.
Signed-off-by: David S. Miller <davem@davemloft.net>
Tim Gardner [Wed, 20 Apr 2011 09:00:49 +0000 (09:00 +0000)]
atl1c: Fix work event interrupt/task races
The mechanism used to initiate work events from the interrupt
handler has a classic read/modify/write race between the interrupt
handler that sets the condition, and the worker task that reads and
clears the condition. Close these races by using atomic
bit fields.
Cc: stable@kernel.org Cc: Jie Yang <jie.yang@atheros.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ivan Vecera [Thu, 21 Apr 2011 00:20:04 +0000 (00:20 +0000)]
be2net: increment work_counter in be_worker
The commit 609ff3b ("be2net: add code to display temperature of ASIC")
adds support to display temperature of ASIC but there is missing
increment of work_counter in be_worker. Because of this 1) the
function be_cmd_get_die_temperature is called every 1 second instead
of every 32 seconds 2) be_cmd_get_die_temperature is called, although
it is not supported. This patch fixes this bug.
Signed-off-by: Ivan Vecera <ivecera@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Thomas Egerer [Wed, 20 Apr 2011 22:56:02 +0000 (22:56 +0000)]
ipv6: Remove hoplimit initialization to -1
The changes introduced with git-commit a02e4b7d ("ipv6: Demark default
hoplimit as zero.") missed to remove the hoplimit initialization. As a
result, ipv6_get_mtu interprets the return value of dst_metric_raw
(-1) as 255 and answers ping6 with this hoplimit. This patche removes
the line such that ping6 is answered with the hoplimit value
configured via sysctl.
Signed-off-by: Thomas Egerer <thomas.egerer@secunet.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:43 +0000 (07:57 +0000)]
tg3: Add additional EEE messaging
This patch adds link messages and an item to the sign-on banner to make
EEE status more visible.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:42 +0000 (07:57 +0000)]
tg3: Add macro for SMDSP toggling
A common AUX CTRL operation in the driver is to enable and disable the
SMDSP. This patch consolidates the code so that the details of the
operation are in one place. This patch also adds code to make sure the
SMDSP is enabled before executing code that relies on it.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:41 +0000 (07:57 +0000)]
tg3: Add write accessor for AUX CTRL phy reg
This patch adds a write accessor for the aux ctrl phy register.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:40 +0000 (07:57 +0000)]
tg3: Add read accessor for AUX CTRL phy reg
This patch adds a read accessor for the aux ctrl register.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:39 +0000 (07:57 +0000)]
tg3: Move phy accessor functions higher
Phy accessor functions should live closer to where the base phy read /
write routines are.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:38 +0000 (07:57 +0000)]
tg3: Only allow phy ioctls while netif_running
When tg3 was new, phy accesses through ioctl were allowable at any time.
Then, the driver started shutting down the phy when the device was
closed. Phy accesses would be allowed when the driver first attached to
the device, but then would be forbidden after the device had been up'd
and down'd. After that, management firmware made it illegal to access
the phy unless the driver "owned" the device. Now that most firmware
is being moved over to the APE, it is less clear when phy accesses are
safe.
While it is possible to attempt to identify these conditions and code
the driver to navigate through the pitfalls, it could be perplexing to
the admin why phy accesses work in some cases and not others. This
patch brings some uniformity to the problem by only allowing phy
accesses while the driver has control of the device.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:37 +0000 (07:57 +0000)]
tg3: Nullify RSS for loopback test
The loopback test assumes all traffic goes to the first rx queue. There
is a 1 in 4 chance this won't be true if RSS is enabled though. This
patch reprograms the RSS indirection table to route all rx packets to
the first queue.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:36 +0000 (07:57 +0000)]
tg3: Adjust rx prod ring bd replenish thresholds
The oldest tg3 devices had large rx producer ring BD caches. Back then,
it made sense to make the BD cache replenish threshold only a function
of the number of rx buffers posted by the driver. Since then, the BD
cache sizes have shrunk to 25% of their original size and, in some
cases, the ring sizes have quadrupled in size. Under such conditions,
static BD cache replenish thresholds no longer match the hardware
constraints.
This patch attempts to factor in the BD cache size into the bd cache
replenish strategy, taking the existing hardware bugs into account.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Matt Carlson [Wed, 20 Apr 2011 07:57:35 +0000 (07:57 +0000)]
tg3: Workaround rx_discards stat bug
The 5717, 5718, 5719 A0, and 5720 A0 has a bug where the rx_discards
statistic counter will increment when dropping unwanted multicast
frames. This patch works around the problem by attempting to
recreate the data using other means. The resulting value will not be
accurate, but it can still serve as a problem indicator.
Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Andrea Galbusera [Thu, 21 Apr 2011 02:21:21 +0000 (02:21 +0000)]
powerpc: Fix multicast problem in fs_enet driver
mac-fec.c was setting individual UDP address registers instead of multicast
group address registers when joining a multicast group.
This prevented from correctly receiving UDP multicast packets.
According to datasheet, replaced hash_table_high and hash_table_low
with grp_hash_table_high and grp_hash_table_low respectively.
Also renamed hash_table_* with grp_hash_table_* in struct fec declaration
for 8xx: these registers are used only for multicast there.
Tested on a MPC5121 based board.
Build tested also against mpc866_ads_defconfig.
Signed-off-by: Andrea Galbusera <gizero@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / mirror_ubuntu-disco-kernel.git / log